To determine whether baseline individual nutritional biomarkers can predict functional response to a multicomponent exercise program in older adults during acute hospitalization.

Secondary analysis of a randomized controlled trial.

Acute Geriatric Unit of a tertiary public hospital in Navarra, Spain.

A total of 303 hospitalized patients aged ≥75 years.

Multicomponent exercise intervention (resistance, balance, and gait training) versus usual care.

Functional response was defined as an increase of ≥1 point in the Short Physical Performance Battery (SPPB) score from admission to discharge. Nutritional status was assessed using baseline serum biomarkers (total cholesterol, albumin, total proteins, and vitamin D) and the Mini Nutritional Assessment (MNA) questionnaire.

In the exercise group, 85.3% of participants were functional responders. Responders had higher baseline total cholesterol levels (148 vs. 126.5 mg/dL; p = 0.01) than non-responders. Adjusted logistic regression confirmed that cholesterol was a significant predictor (OR 1.018 per mg/dL; 95% CI 1.00–1.03; p = 0.014). In the overall cohort, a significant interaction between baseline cholesterol and exercise intervention was identified (interaction OR 1.025 per mg/dL; 95% CI 1.008–1.041; p = 0.004). Also, a significant interaction between total serum protein levels and exercise was observed (interaction OR 3.17 per g/dL; 95% CI 1.16–8.66; p = 0.024).

Baseline total cholesterol, which may reflect metabolic reserve, predicts and modulates the functional response to in-hospital exercise. These findings, together with the additional interaction observed for total serum proteins, suggest that metabolic reserves enhance the capacity to respond to rehabilitation stimuli. Nutritional assessment should be integrated into rehabilitation strategies to identify older adults with physiological resilience who can benefit from such programs.